When controlling the internal combustion engine in the neutral position, as fast as possible a response to a deviation in the rotational speed from the desired setpoint value is crucial for the quality of idle running. This is particularly the case if an unknown load is applied, and for start-up.
Not only the measured rotational speed but also the rotational speed gradient calculated therefrom is used for the control. However, the problem with using the rotational speed gradient for control is that the rotational speed gradient—even under stationary conditions—is never really constant. This is because of measuring errors which result from mechanical and electrical tolerances when recording the measured values and fluctuations in synchronism of the internal combustion engine. Hence until now only the filtered rotational speed gradient has been used for control, with short-term fluctuations being filtered out.